How Many B2B SaaS Companies Are There in 2026?💻

There is no government register of SaaS companies. No trade body maintains one. Every figure in circulation traces back to a commercial database counting something slightly different: funded companies, product listings, vendor profiles, or revenue-verified businesses. Those definitions produce answers ranging from 63 to more than 200,000.

The number a marketer picks does not describe the market. It describes which database was consulted.

The competing answers, side by side

Same question. A spread of more than 2,700x.

Figure 1. Published counts of SaaS companies, 2026. Logarithmic scale.

The most-repeated figure collapses on its own source page

The number that dominates search results is 30,800 globally with 17,000 in the United States. It appears in nearly every SaaS statistics roundup published for 2026. Its most-cited origin is Ascendix, which attributes the total to a Statista dataset.

That page contradicts itself repeatedly:

  • It opens by placing 30,800 companies at the start of 2025, then attributes the same 30,800 to the end of 2024.
  • Three sentences later it states that by the end of 2024 there were around 72,000 SaaS companies in operation, and that including artificial intelligence companies could raise the figure to 175,000. Three incompatible totals in one section.
  • The United Kingdom appears as approximately 1,500 companies in one paragraph and 2,000 in another.
  • Segment breakdowns list over 24,000 productivity and collaboration tools, about 17,000 in customer service, 15,000 in marketing, nearly 14,000 in e-commerce, 12,000 in data and analytics, and 11,500 in sales. That sums to more than 93,000, roughly triple the stated global total.
  • It claims around 450 new SaaS companies were created globally in 2023 and approximately 500 as of mid-2024. Moving from 25,000 companies in 2021 to 30,800 in 2024 requires roughly 1,900 net additions per year, four times the stated rate.

Every roundup republishing 30,800 as a 2026 figure inherited a 2024 number from a page that cannot reconcile its own arithmetic.

Two databases, two incompatible world maps

Backlinko sources its count to Latka: approximately 42,000 SaaS companies worldwide, with 12,400 in the United States, 1,700 in the United Kingdom, 1,100 each in Canada and India, and 916 in Germany.

Set that against the Statista lineage of 30,800 worldwide and 17,000 in the United States.

One dataset puts the United States at 30% of global SaaS. The other puts it at 55%. Both cannot describe the same market. Both are published as fact in 2026 roundups, frequently within the same article.

Figure 2. US share of global SaaS under two competing datasets.

The Backlinko page carries a 2026 title and a December 2025 update stamp, but its supporting data points to Gartner forecasts from 2024 and Dealroom funding figures from the first half of 2024. The year in the headline is a publishing decision, not a data refresh.

Companies, products, vendors and listings are not the same unit

The largest single driver of the spread is unit confusion. Directories count products. Databases count companies. Marketplaces count vendor accounts. One company can hold several product listings across several platforms.

The G2 awards page is the clearest illustration. Out of 151,264 total products on the platform, only 4,692 were eligible for the 2026 Best Software Awards.

Figure 3. G2 listings versus trading businesses.

Eligibility requires a modest threshold of recent verified reviews. Just 3.1% of listed products clear it. The remaining 96.9% are profiles, many of them dormant, abandoned, or never activated.

That gap is the real finding. A directory count answers how many software profiles exist, not how many software businesses are trading. Those are different questions with different answers, and market-sizing exercises routinely conflate them.

The concentration is now structural. G2 agreed in January 2026 to acquire Capterra, Software Advice and GetApp from Gartner, bringing four of the leading B2B software review platforms under one owner. Three of the independent counts listed above now sit inside a single company.

One category alone breaks the global total

Scott Brinker and Frans Riemersma have enumerated the marketing technology landscape annually since 2011 using a published methodology. This is a genuine bottom-up count, not a database export.

The 2026 landscape contains 15,505 products, up 121 from 15,384 the previous year, growth of 0.79%.

Marketing technology is one category among dozens. Human resources technology, financial technology, developer tools, cybersecurity, health technology, legal technology, supply chain and vertical SaaS each support ecosystems of comparable density.

If a single category holds 15,505 products, a global total of 30,800 SaaS companies would make marketing technology half of all SaaS on earth. That is not plausible. The most-cited number in the category fails a one-line sanity check.

The strictest definition produces 63

At the other end sits the SaaS Capital Index, which applies a rigorous definition to public markets. It contains 63 publicly traded companies as of 30 June 2026, all listed on United States exchanges and judged to be valued primarily on B2B recurring software revenue.

The exclusions reveal how much definitional work the phrase “B2B SaaS company” is doing:

  • Companies serving B2C customers are excluded, along with B2B companies whose annual revenue per customer falls below $500. Box, Dropbox, Carbonite and 2U are named examples.
  • Businesses with significant perpetual licence, maintenance, hardware or telecom revenue are excluded. Oracle, Microsoft, Square and Twilio are named examples.
  • Software consolidators such as Constellation Software are excluded for business model reasons.

Under a defensible, published standard, Microsoft and Oracle are not B2B SaaS companies. Under the Ascendix framing, Microsoft is the largest one in the world. Neither is wrong. They answer different questions using the same words.

Government statistics have no SaaS category at all

This is why no official count exists.

The closest United States industry code is NAICS 513210, Software Publishers. Its index entries are built around packaged software: packaged applications, packaged operating systems, packaged utilities, packaged compilers. There is no entry for SaaS, cloud software, or subscription software as a business model.

The cross-references then split the SaaS value chain three ways. Hosting and infrastructure support provided to software publishers is classified under 518210, Computing Infrastructure Providers, Data Processing, Web Hosting and Related Services. Custom software built for specific users falls under 541511, Custom Computer Programming Services.

The companies grouped inside 513210 by annual sales include Microsoft, Oracle, Salesforce, Adobe, Activision Blizzard, Zillow Group, Block and Global Payments. A CRM platform, a video game publisher, a real estate portal and two payments processors share one code.

No statistical agency counts SaaS companies because SaaS is not an industry in the classification system. It is a delivery and pricing model that cuts across many. Every number in the market exists because commercial vendors filled that vacuum with their own inclusion rules.

Nobody is isolating B2B

The question asked here concerns B2B SaaS specifically. Almost none of the published figures attempt that split.

The Ascendix FAQ places approximately 17,000 B2B SaaS companies globally and, two answers later, approximately 17,000 SaaS companies in the United States. Identical figure, two different denominators, one page. The B2B number is the United States total relabelled.

The only sources applying a genuine B2B filter are those with commercial reasons to be strict. The SaaS Capital index of 63 is one. Its fifteenth annual survey, completed in March 2026, drew responses from more than 1,000 private B2B SaaS companies, which is a sampling frame rather than a population count.

A stock number without a flow number is not a market size

The martech data exposes what every static count omits. In 2026, 1,488 products were added to the landscape and 1,367 were removed, so additions and removals nearly cancelled out.

Figure 4. Martech products added versus removed, 2025 and 2026.

The composition of those exits matters more than the net figure. The largest cohort of departures came from the 2010 to 2019 SaaS wave, accounting for 51.7% of removals. By revenue, 45.5% of removed products sat in the $1M to $10M range. By headcount, 41.2% had one to ten employees and 38.7% had eleven to fifty.

Nearly 9% of a mature category turned over in twelve months, concentrated in exactly the segment most B2B SaaS companies compete in. A headline count reporting only the standing total conceals that churn entirely.

What to use instead

For any B2B SaaS team doing market sizing, competitive analysis or addressable market modelling, the global company count is the wrong input. Three substitutes hold up:

Count the category, not the market. Enumerated category landscapes carry stated methodology and year-over-year comparability. The chiefmartec landscape is the model. Find or build the equivalent for the relevant vertical.

Count activity, not existence. The G2 ratio of 4,692 review-active products against 151,264 listings is a better proxy for live competitors than any company total. A competitor with no customers is not a competitor.

Count the flow. Additions and removals reveal category maturity. A category adding 1,488 and removing 1,367 is consolidating. One adding 2,489 and removing 1,211, as martech did the previous year, is still expanding. Positioning decisions differ sharply between the two.

For content and messaging, the practical implication is narrower. Buyers do not care how many SaaS companies exist. They care how many solve their problem, which is a category question answerable with an enumerated landscape and a review-activity filter. That is also the count determining how hard it is to rank in search and to be cited in AI answers, which is where B2B SaaS demand is increasingly decided.

Methodology

Every figure was verified by direct retrieval from the publishing source. Contradictions were preserved rather than reconciled, because the contradictions are the finding.

Figures reported by CloudZero covering Capterra, Vainu and SaaSHub were not independently confirmed at the originating platform and should be treated as directional. The Statista dataset sits behind a paywall and was assessed through the citing page, which is itself part of the problem this article documents.

Related benchmark data: B2B Customer Retention Statistics.

Verification logbook